The diffuse injury to peripheral nerves (diabetic neuropathy) is exceptionally common in type 1 diabetes, but there is a lack of an objective surrogate marker to identify early subclinical stages when treatments might be most effective, prior to late-stage progression to troublesome and costly foot infection, ulceration, and limb amputation. In contrast to the ability to objectively measure disease-specific surrogate markers for retinopathy and nephropathy, this lack of a diabetic neuropathy surrogate marker has seriously impeded the development of specific interventions in clinical research trials. Representing 5 independent research groups that have together created a consortium of investigators dedicated to the development of a surrogate marker for early diabetic neuropathy, we have focused on using the eye as a window to non-invasively image by a method of in-vivo corneal confocal microscopy (CCM) the small nerve fibres that innervate the cornea. We have demonstrated that changes in these nerve fibre endings occur early in the development of neuropathy, reflect well the changes seen in other peripheral nerves by invasive skin biopsy evaluation, and that their measurement is feasible and reproducible. As a multinational consortium, we have the benefit in this proposal of pooling multiple cohorts to apply the most valid study methods in biomarker development. First, we aim to determine in the analysis of an existing pooled dataset of 516 type 1 and 524 type 2 diabetes subjects the exact levels of CCM measurement that can identify the presence of diabetic neuropathy. Secondly, we propose over three years to re- examine at least 70% of this cohort, which will provide 5- to 7-year follow-up data to determine which type and level of CCM measurement can predict the future onset of neuropathy, as well as its progression in those who had neuropathy at baseline. Finally, we will evaluate the role of time- and cost-saving automated image analysis software. By virtue of large sample size from data pooling, we are uniquely afforded the methodological power to confirm our objectives by way of separate derivation and validation analysis sets. Through a unique and unprecedented multinational pooled dataset approach for diabetic neuropathy, this work will derive and validate specific CCM parameter thresholds for the identification of neuropathy, and - more importantly - the identification of individuals at future risk. These results will permit application in clinical practice and intervention trials for neuropathy risk stratification. Evaluaion of automated image analysis will influence likelihood of successful knowledge translation of this surrogate biomarker into clinical practice - in which the procedure could be harmonized with annual retinal examinations - and into intervention trials.